Benford's law only works in multiple ranges of magnitude, where some counties are hundreds of votes, thousands of votes, tens of thousands of votes, etc....

An easy example is between 10-199, you're naturally going to have more numbers starting with 1 because 100-199 all start with one. 10-299 will have more 1s and 2s because 200-299 all start with 2, all the way to 999 till you have 9s to equal out, then from 1000-1999, they all start with ones, so there will be more 1s again.

But if the orders of magnitude are all the same, it's different. Say it was 100 counties with populations between 100 - 999, with most counties being in the middle. If you charted a graph of first number of the population of county, you'd get a "Normal" bell curve like graph, with 4 and 5 being the most common and and 1 and 9 being the least. If the populations were random between 100-999, it would be a flat line.

In Nebraska, I'm guessing many of these counties are 4 digits, so it would skew the graph.